| 有限混合泊松分布参数优化的改进EM算法 |
| |
| 引用本文： | 冯杭,王胜兵.有限混合泊松分布参数优化的改进EM算法[J].兵工自动化,2017,36(1):80-82. |
| |
| 作者姓名： | 冯杭 王胜兵 |
| |
| 作者单位： | 海军工程大学理学院,武汉,430033;海军工程大学理学院,武汉,430033 |
| |
| 摘 要： | 为解决 EM 算法易于陷入局部最优的问题，提出一种有限混合泊松分布参数优化的改进 EM 算法。建立混合泊松分布模型，利用 EM 算法求解模型中的参数向量，并结合算例指出该算法对初值敏感的缺陷，引入智能优化算法对算例进行改进。分析结果表明：使用优化算法后，对数似然度的增大十分明显，增加了求得最优解的概率。
|
| 关 键 词： | 泊松混合分布 最大似然估计 期望最大化 智能优化算法 |
| 收稿时间： | 2017/2/10 0:00:00 |
| 修稿时间： | 2016/5/27 0:00:00 |
Improved EM Algorithm of Limited Mixed Poisson
Distribution Parameters Optimization |
| |
| Feng Hang.Improved EM Algorithm of Limited Mixed Poisson
Distribution Parameters Optimization[J].Ordnance Industry Automation,2017,36(1):80-82. |
| |
| Authors: | Feng Hang |
| |
| Affiliation: | College of Science, Naval University of Engineering, Wuhan 430033, China |
| |
| Abstract: | For solving problem that EM algorithm is easy involved in local optimization, put forward improved EM algorithm of limited mixed Poisson distribution parameters optimization. Establish mixed Poisson distribution model, use EM algorithm to solve model parameter vector, combine example to point out the defect of algorithm in initial value sensitivity, use intelligent optimization algorithm to improve example. The analysis results show that: by using optimization algorithm, the log likelihood ratio is increased significantly, and greatly improve probability of optimization solution. |
| |
| Keywords: | poisson mixed distribution maximum likehood estimation (MLE) expectation-maximization (EM) intelligent optimization algorithm |
| 本文献已被 万方数据 等数据库收录！ |
| 点击此处可从《兵工自动化》浏览原始摘要信息 |
|
点击此处可从《兵工自动化》下载全文 |
|
